Rotated XZ and YZ arcs in emc 2.4 are displayed incorrectly by the
preview plot in axis.  To see what I mean, MDI G10 L2 P1 R30 and then
load 3dtest.ngc.

They are now correct in master (or mostly right; I keep finding new bugs
when I look) but only after a fairly comprehensive rewrite, using C++
code instead of Python code for a certain workhorse function.  (Thanks
to Alexey Starikovskiy for the original impetus and his help in this
matter, by the way)

It seems to me that there are three things we might choose to do about
this in 2.4:
 * nothing; document that it is broken and live with it
 * backport the changes from emc2.4
 * write a new Python arc implementation that properly rotates these
   arcs

These each have problems.  In the first case, the software is broken.
In the second case, we are putting fairly radical changes into the
"stable" branch (moving methods from Python to C seems radical to me).
In the third case, we're taking a separate stab at solving a complex
problem, and risk making arcs in 2.4 slower than they are now (speed was
the motivation for the original changes in master).

I think that the second alternative is probably the best of the three.
For that reason, I've backported and lightly tested the arcs code from
master to v2.4_branch.  There were fairly extensive conflicts, because
many of the patched locations were changed by the "glrefactor" change
on master.  I've put the resulting patches online:
http://emergent.unpy.net/files/sandbox/axis-c-arcs.mbox

Jeff Epler (15):
      reduce python overhead in arcs
      Convert arcs to segments in C
      reduce minimum number of arc segments
      allow tuning of arc subdivision at this layer
      allow inifile to set arc subdivision
      reduce default number of segments
      document new inifile item
      make it build with python2.4
      simplify populating n[]
      calculate tx, ty without requiring rad
      this reformulation allows g++ to see that plane is initialized
      fix helical arcs
      the machine coordinate system does not rotate
      make a certain "full circle" arc appear
      another fix for helical arcs

 docs/src/config/ini_config.lyx        |   28 ++++++
 lib/python/rs274/glcanon.py           |   14 ++-
 lib/python/rs274/interpret.py         |   63 +------------
 src/emc/rs274ngc/gcodemodule.cc       |  158 ++++++++++++++++++++++++++++++++-
 src/emc/usr_intf/axis/scripts/axis.py |   15 ++--
 5 files changed, 207 insertions(+), 71 deletions(-)

If you have an interest in this issue, please build with these changes
and let the mailing list know if you spot any problems.  If you think
one of the other alternatives should be pursued instead, please make
that case.

Jeff

------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Emc-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-developers

Reply via email to